Abstract:The smart grid (SG) is a modern power system that leverages digital technologies to enhance the generation, delivery, and consumption of electric power. Reinforcement learning (RL) plays an important role in SG by helping make smart decisions. However, RL faces challenges such as sparse rewards, poor generalization, and difficulty in representing and interpreting regulatory constraints. Large language models (LLMs) offer new opportunities to address these challenges by understanding natural language, leveraging external knowledge, and enhancing reasoning. This paper presents a comprehensive review of LLM-enhanced RL for SG. It first analyzes the key challenges of RL and introduces how LLMs help enhance the performance, including a taxonomy based on the integration patterns and functions. It then reviews applications of LLM-enhanced RL for SG, with a focus on energy management, operational control, electricity market, and hardware design. Finally, it discusses research challenges and future directions of LLM-enhanced RL for SG. This paper aims to provide a clear framework for researchers to apply LLMs in RL and to promote the application of RL for SG.

Abstract:’With the global energy transition towards more decentralized and sustainable systems, peer-to-peer (P2P) energy trading has emerged as a prominent energy exchange model, attracting increasing attention from both academia and industry. P2P energy trading not only facilitates the efficient utilization of distributed energy resources (DERs) but also enables autonomous energy transactions. This paper provides a comprehensive review on P2P energy trading, starting with an explanation of the P2P energy network, P2P market structures, and blockchain in P2P energy trading. The main body reviews existing research from six key research domains: trading platform, market clearing method, security, uncertainty, modeling of users behavior, and experimental validation. Subsequently, the paper discusses key factors in deployment of P2P energy trading systems. Finally, opportunities and challenges in the future are outlined, and conclusions are drawn.

Abstract:Assessments of equivalent inertia and primary frequency regulation (PFR) are critical for evaluating frequency stability in modern power systems. Traditional methods typically aggregate the contributions of all generation units or rely on system identification during disturbances. However, these methods may yield inaccurate or incomplete results in systems with a high penetration of inverter-based resources (IBRs), as IBRs may reach their power reserve limits (PRLs) and fail to provide the expected frequency support. To overcome this limitation, this paper proposes an assessment method of equivalent inertia and PFR of power systems considering PRLs of IBRs. In addition, dead zones and control delays are incorporated to further enhance the accuracy of the assessment. Based on the assessment results, an optimal parameter-setting approach for equivalent inertia and PFR is then developed. Notably, the theoretical optimal settings are generally difficult to achieve in practice, implying that conventional parameter settings may lead to significantly lower equivalent inertia than expected. The effectiveness of the proposed method is validated through time-domain simulations.

Abstract:This paper presents a frequency constrained economic dispatch (FCED) method designed to prevent frequency violations under N – 1 contingencies, encompassing both generator outages and high-voltage direct current (HVDC) infeed losses. Firstly, the coordinated frequency dynamics of synchronous generators, wind turbines, and HVDC lines are modeled, and the resulting frequency constraints are derived. To address the challenging frequency nadir (FN) constraint, a physics-informed piecewise linearization (PI-PWL) model is introduced to learn the FN security boundary. The formulation of the proposed FCED method is then presented, incorporating frequency constraints under N – 1 contingencies. Additionally, an error-aware constraint cut (EACC) is proposed to eliminate errors introduced by data-driven methods and accelerate solution convergence. Case studies on the modified IEEE 39-bus and IEEE 118-bus test systems validate the effectiveness of the proposed FCED method. Results indicate that the proposed PI-PWL model more accurately captures the FN security boundary than the traditional PWL (T-PWL) model, particularly in scenarios with limited unstable samples. Moreover, the proposed FCED method ensures frequency security under N – 1 contingencies, while the EACC effectively eliminates approximation errors and accelerates convergence compared with fixed-step constraint cuts.

Abstract:Wind turbine (WT) is required to support grid frequency in some cases usually by decelerating the rotor speed. However, the available adjustment capacity is limited and short-lasting, requiring reserve power for more capacity. For wind farms (WFs), reserve power dispatching is important both at steady state and during dynamic grid frequency support process, as it influences the lifespan and cost of WFs. Moreover, power dispatching at steady state and during dynamic grid frequency support process may have a mutual impact on their performance. For this regard, this paper proposes a hierarchical reserve power dispatching (HRPD) strategy of WFs for grid frequency support. The hierarchical concept is reflected on the WF and the WT control layers spatially, as well as on the steady-state layer and the dynamic grid frequency support layer temporally. First, the central controller of WF determines the reserve power command of each WT at the steady-state layer according to the total reserve power requirement, where the total fatigue of WFs and the fatigue distribution among WTs are considered to reduce the operating fatigue of the WF and maintain the consistency of WTs. Meanwhile, the WT controller assesses the real-time grid frequency support capability of each WT at the dynamic grid frequency support layer, according to the available kinetic energy of rotor and the real-time pitch angle to optimize the utilization of regulation capacity while maintaining the performance of grid frequency support. Based on this, the central controller adjusts and assigns a dynamic frequency droop coefficient to each WT. Case studies validate the performance of the proposed HRPD strategy in total fatigue reduction, fatigue distribution restriction, and grid frequency support.

Abstract:Forced oscillation source localization (FOSL) using synchrophasor measurements is critical for mitigating forced oscillations (FOs) in power systems. However, the performance of the conventional dissipating energy flow (DEF) is significantly affected by measurement noise and other irrelevant modal components. To address this challenge, an optimal subspace-enhanced (Os-enhanced) dynamic mode decomposition (DMD)-assisted DEF-based FOSL is proposed in this paper using synchrophasor measurement. First, Os-enhanced DMD is employed to decompose multi-channel measurements into time-domain responses of individual modes. Then, the time-domain components associated with FO are distinguished from the decomposed modes based on the rate of change of modal energy, and are subsequently used to calculate the DEF at all generator buses. Furthermore, the rate of forced energy is introduced as an indicator of energy flow direction to localize the FO source. The performance of the Os-enhanced DMD-assisted DEF is evaluated by the simulated data from IEEE 16-machine 5-area system and field phasor measurement unit (PMU) measurement data from Independent System Operator New England (ISO-NE). Comparative results demonstrate that the Os-enhanced DMD-assisted DEF achieves improved accuracy and robustness for localizing FO sources under noisy conditions.

Abstract:Deep reinforcement learning (DRL) has been recognized as a promising alternative for emergency controls recently, for its rapid decision-making and strong policy searching capabilities. However, when applied to complex large-scale power systems, two prominent challenges emerge>① high-dimensional discrete-continuous hybrid action space poses great adaptation difficulty for regular DRL methods, as they are designed primarily for purely discrete or continuous action spaces; and ② unguided exploration within vast action spaces leads to inefficient convergence performance. To mitigate these limitations, this paper develops a knowledge-guided hybrid DRL-based method for transient stability emergency control. The proposed method adopts a novel hybrid policy to represent the hybrid action domain, and employs a novel derivative-free DRL for optimization, which eliminates the training instability issues associated with jointly optimizing gradients for different action types. Then, the prior domain knowledge is utilized to formulate mask rules and incorporate them into DRL training through trainable action mask (TAM) technique to guide exploration. Moreover, the proposed method is further integrated into a hierarchical DRL framework to alleviate computational complexity and enhance scalability. Case studies on the IEEE 300-bus system show that the proposed method improves solution quality by 52.9% and 42.8% compared with purely discrete and conventional hybrid action DRL methods, and achieves an 80.0% improvement in convergence efficiency compared with the method without knowledge guidance.

Abstract:The current saturation challenges the transient stability of grid-forming (GFM) inverter-based resources (IBRs) during large disturbances by introducing the current-saturated stable equilibrium point (CS-SEP). Convergence into the post-fault CS-SEP is an undesired post-disturbance scenario. The angle of the saturated current significantly affects the post-disturbance trajectory. Moreover, the voltage immediately after the converter returns to the normal operation mode depends on this angle. If the difference between this voltage and the reference voltage is large, huge power oscillations occur superimposed on the second-order power swing. The voltage-error-induced power oscillation appears as a disturbance on the active power control loop and affects transient stability. In this paper, the angle of the saturated current is adjusted to minimize the voltage error immediately after the GFM IBR returns to the normal operation mode and prevent the convergence into CS-SEP. To do so, a closed-form expression for the angle of the saturated current at which the voltage transient is minimized is derived in the first stage. Then, an optimization is formulated to minimize the deviation of the angle of the saturated current from the angle calculated in the first stage while ensuring that CS-SEP convergence is avoided. This optimal control method, which enhances transient stability by adaptively tuning the angle of the saturated current, is validated through simulation.

Abstract:In 100% power electronics-based power system, small-signal synchronous instability can be trigged by interaction among distributed grid-forming converters (GFMCs). A novel small-signal synchronous stability analysis method is proposed. Firstly, a generic small-signal model for 100% power electronics-based power system is established. Then, a framework based on a dominated synchronization control loop is proposed, which can explicitly identify two main interaction paths stemming from voltage control and reactive power-voltage control. Furthermore, the two interaction paths have been simplified to first-order transfer functions through model reduction based on dominated modes. By integrating the selected synchronization control loop, a reduced second-order model that can provide clear physical insight into small-signal synchronous stability is derived. Finally, experimental results from an RT-LAB hardware-in-the-loop platform confirm the effectiveness of the proposed method.

Abstract:The stability of microgrid (MG) is affected by the interaction between parallel inverter-based resources (IBRs), especially when grid-forming (GFM) control is implemented in islanded MG to provide voltage support. In this paper, the robust stability analysis using structured singular value (SSV) approach and design of parallel GFM converters in islanded MG by structured μ synthesis are presented, catering to its multi-input multi-output (MIMO) nature. Specifically, a modular state-space model of parallel GFM converters in islanded MG is derived with component connection method (CCM). Model uncertainty is captured against the nominal model, delivering a certain degree of robustness for stability analysis, in the face of potential parametric uncertainty and resonance effects induced by parallel GFM converters. Moreover, the effects of system parameters on robust stability are comprehensively analyzed, and a robust design is proposed for virtual impedance (VI) tuning. The accuracy of modelling and robust stability analysis, as well as the effectiveness of the robust design is validated in MATLAB/Simulink/PLECS and real-time digital simulator (RTDS).

Abstract:The grid-forming (GFM) energy storage-based static synchronous compensator (ES-STATCOM) operates as a voltage source and offers virtual inertia and damping. These capabilities make it more suitable for the modern power system with high renewable energy penetration than grid-following (GFL) static synchronous compensator (STATCOM). However, it is prone to overcurrent during grid faults. Physically increasing the overload capacity of the device is not a cost-effective solution. In this paper, a current limiting strategy based on virtual sequence impedance for the GFM ES-STATCOM under asymmetrical faults is proposed. It enables the GFM ES-STATCOM to maintain its GFM control mode during transients. The control for activation and deactivation of current limiting is first designed. Then, the transient virtual sequence impedance control is developed to limit the fault current amplitude, along with the instantaneous overcurrent peak suppression method. In addition, key control parameters are designed through analysis of fault current characteristics. Finally, the proposed strategy is validated by PSCAD/EMTDC simulation under asymmetrical faults.

Abstract:With the rapid development of the electricity market and the increasingly stringent operational standards, the longstanding infeasible network-constrained unit commitment (NCUC) issue in market operations is calling for effective and systematic methods that can analyze and repair infeasible NCUC models. Studies show that a set of irreducible infeasible subsets (IISs) may lead to infeasibility in a model. The identification and repair of IISs can correct the original infeasible model, while traditional filtering algorithms for IIS location often exhibit computational inefficiency. A fast infeasibility analysis framework of NCUC and a relaxation filtering algorithm are proposed in this paper to analyze infeasible NCUC models more comprehensively and effectively. The infeasibility analysis framework employs subsystem-based IIS analysis method, generating and analyzing simplified models to repair the infeasible NCUC. The relaxation filtering algorithm can achieve greater effectiveness of IIS identification in the holistic NCUC compared with traditional general filtering algorithms. The case studies show that the proposed framework and algorithm can analyze and repair infeasible NCUC models effectively, reducing the analytical time by at least one order of magnitude compared with that of traditional general methods.

Abstract:Conditional transmission section limits (C-TSLs) depend on the actual operational state of the power grid, such as commitment status and reserve capacity, exacerbating the computational complexity of the unit commitment problem. To overcome this complexity, this paper proposes a data-model hybrid-driven approach for solving unit commitment problems. Furthermore, a unit commitment prediction algorithm is proposed to reduce the problem scale by fixing a subset of unit commitment variables, thereby eliminating inactive C-TSL intervals. Specifically, the proposed algorithm incorporates a convolutional attention module to perform deep feature extraction and enhancement on time series data, complemented by a multi-head cross-attention mechanism designed to synthesize temporal and non-temporal features. The introduction of attention mechanisms enhances the prediction performance of the proposed algorithm. Numerical results demonstrate that the proposed approach significantly reduces the number of variables and constraints in the unit commitment model, thereby expediting the solution process while maintaining high solution accuracy.

Abstract:This paper addresses the increasingly tight energy coupling in urban energy systems, which are composed of the power distribution network (PDN), the transportation network (TN), and the gas distribution network (GDN). A non-cooperative game optimization model for the urban energy system is developed, to achieve coordinated optimization among these three different energy networks operated by independent stakeholders. To avoid excessive sharing of private information, an inner-outer iterative method is further proposed to obtain the Nash equilibrium and enhance solving efficiency. In the proposed method, each energy network is iteratively optimized by exchanging only the price and load information, and the network parameters are not disclosed. In the inner layer, the PDN-TN coupled subsystem and the PDN-GDN coupled subsystem are solved separately. In the outer layer, interactions between charging loads and gas prices are coordinated through the PDN, thereby enabling cross-network coordination of urban energy systems. Case studies demonstrate that the convergence speed of the PDN-TN coupled subsystem is significantly improved by dynamically adjusting the electricity price based on sensitivity coefficients. Balanced resource allocation of the urban energy system is achieved through the non-cooperative game, while autonomy of all stakeholders is preserved. Energy procurement costs are reduced by up to 25.56% compared with that of the PDN-TN coupled subsystem.

Abstract:The increasing interdependence between transportation networks (TNs) and distribution networks (DNs) presents significant challenges for the analysis and risk assessment of coupled networks. This paper presents a risk assessment framework of coupled TN and DN (TDN) considering charging power restriction and uncertainty factors, simulating operations through a coordinated optimization model. The model incorporates charging power restrictions for fast charging stations (FCSs). To enhance computation efficiency, nonlinear constraints are linearized, and an accuracy-aware adaptive piecewise linearization method is utilized. Random variables within the coupled TDN are employed as inputs to develop a surrogate model based on the Gaussian process regression method, complemented by suitable kernel functions designed for the scenarios with numerous discrete categorical variables. A global sensitivity analysis is conducted to identify the continuous and discrete uncertainty factors that significantly impact system performance. The proposed risk assessment framework serves as a valuable tool for evaluating the coupled TDN and facilitating the identification of continuous and discrete uncertainty factors affecting network operation while revealing risks associated with failures in the operation of coupled TDN.

Abstract:Cybersecurity in power systems with distributed energy resources (DERs) has become a serious challenge. This paper proposes a critical boundary index (CBI)-based preventive and post-detection defense approach against false data injection attacks (FDIAs) in power systems with DERs. The proposed approach utilizes the inherent attack resistance of CBI, a voltage stability index previously developed by the authors, to both preventively limit the impact of attacks and provide high-sensitivity post-attack detection of data manipulation. This paper quantitatively evaluates voltage stability monitoring approaches against FDIAs, comparing the resilience of CBI with conventional indices. Verification is conducted using IEEE 5-bus and 118-bus test systems with two types of FDIA scenarios: detection-avoidance type and misdirection type. The results demonstrate that when CBI is employed for monitoring, the achievable false data injection by attackers aiming to cause voltage drops is significantly reduced compared with conventional indices, maintaining system voltage within stable ranges. Additionally, CBI shows 1.5-2 times higher detection sensitivity compared with conventional indices. Notably, CBI demonstrates effective detection capability even against sophisticated attacks that falsely show improvement in voltage stability indices, which are difficult to detect using conventional approaches. These results confirm that the proposed approach effectively constrains attacker capabilities while providing enhanced detection sensitivity. Based on these findings, the proposed approach is shown to provide high sensitivity to even minor data tampering, offering a multi-layered defense combining prevention and detection for power systems with DERs.

Abstract:The displacement of synchronous generators by converter-interfaced renewable energy sources requires wind farms to provide inertia, damping, and voltage support, particularly in increasingly weak grids. Based on classical frequency-domain loop-shaping techniques, this paper presents a coordinated multi-loop control design methodology of virtual synchronous machine (VSM)-controlled doubly-fed induction generators (DFIGs) in a wind farm to tackle the intra-machine controller interactions. Starting from an initial tuning and a full small-signal linearisation, every local open loop is redesigned to meet explicit phase margin targets through a single and prioritised iteration. The resulting controllers achieve step responses and stability margins close to those programmed at the design stage, despite the cross-coupling between control loops. Results can be improved further if a few more design iterations are carried out. Since the controller synthesis relies exclusively on classical loop-shaping tools available in commercial simulation software, it is directly applicable to industrial-scale projects.

Abstract:Wind turbine (WT) fault diagnosis using supervisory control and data acquisition (SCADA) data is challenged by severe class imbalance, which deteriorates the recognition of minority class. To address this issue, a spatiotemporal feature extraction framework from integrated multivariate time series (IMTS) is proposed, with coordinated designs in data representation, model architecture, and loss regularization. First, a dual-label IMTS is constructed to fuse multi-source SCADA data via sliding windows. Second, a spatiotemporal feature learning architecture is developed by integrating a multi-scale convolutional neural network (MCNN) and a stacked long short-term memory (LSTM) network and designing the MCNN-LSTM model. Finally, an improved weighted cross-entropy loss is proposed, which incorporates kernel density estimation-based sample probabilities as regularization, thereby directing model attention toward minority and hard-to-classify samples. Experimental results on a real SCADA dataset demonstrate that the proposed MCNN-LSTM model achieves a Macro-recall score of 0.943 and a G-mean score of 0.942, outperforming comparison models in comprehensive performance.

Abstract:Holomorphic embedding (HE) method has gained increasing attention in power system analysis because of its robustness. This paper further extends the HE theories into the power flow calculation in hybrid AC/DC active distribution networks (ADNs). A flexible holomorphic embedding solution (FHES) method is proposed for the power flow calculation in AC/DC ADNs in the grid-connected and islanded modes. This paper also develops the HE models of voltage source converters (VSCs) with active and reactive power control modes. The HE models of loads in AC/DC ADNs are also formulated by taking full account of static voltage and frequency characteristics. Furthermore, the HE models of the distributed generators (DGs) under various control modes are developed. A sequential iterative method is employed to solve the developed HE models of AC/DC ADNs in the grid-connected and islanded modes. The performance of the proposed FHES method is evaluated using the modified IEEE 33-node and IEEE 123-node AC/DC ADNs. Finally, the testing results validate the accuracy and robustness of the proposed FHES method.

Abstract:The large-scale integration of photovoltaic (PV) generation poses serious challenges of voltage violations and instability risks to active distribution networks (ADNs). To address these issues, this paper proposes a topology-switching soft open point (TS-SOP) assisted real-time cooperative operation framework of ADNs. The proposed framework ensures voltage stability by explicitly incorporating voltage stability margin (VSM) constraints. A TS-SOP model is developed that enhances power flow controllability and reduces standby losses through switchable feeder interconnections and flexible converter operating modes. The proposed framework coordinates day-ahead pre-dispatch, intra-day dynamic correction, and real-time voltage regulation, featuring a two-stage volt/var control. The first stage employs model predictive control to mitigate PV-driven fluctuations, while the second stage embeds VSM constraints into droop curve parameter optimization, ensuring that local droop control actions respect the system stability limits. Simulations on a modified IEEE 33-node system demonstrates that, compared with conventional fixed-topology SOP and droop-only schemes, the proposed framework not only reduces network losses and voltage violations, but also ensures VSM compliance, effectively coordinating fast voltage regulation with system-wide stability while improving the operating economy.

Abstract:Real-time state estimation underpins advanced applications in distribution networks by revealing system operating conditions with high fidelity. Intelligent measurement terminals increasingly combine heterogeneous measurements, including micro-phasor measurement units (μPMUs), remote terminal units (RTUs), and data transmission units (DTUs). Their complementary strengths can overcome traditional limits in time resolution and accuracy. However, heterogeneous formats, asynchronous updates, and complex temporal dynamics challenge conventional estimators. This paper proposes a unified real-time forecasting-aided state estimation (FASE) framework for distribution networks with multi-source measurement data fusion that fuses μPMU, RTU, and DTU data. A long short-term memory (LSTM)-based data imputation method is designed to effectively align delayed RTU and DTU updates with μPMU sampling, markedly reducing imputation errors versus linear and historical-average baselines. Unified measurement equations are formulated for all devices. An improved cubature Kalman filter (CKF) with adaptive robust weighting is adopted to enhance numerical stability and outlier resilience. To capture multimodal operating regimes driven by variable distributed energy resources, a Gaussian mixture model (GMM)-based approach is integrated into the distribution network state transition. Validated on the IEEE 33-bus and 118-bus systems, the proposed FASE framework achieves higher estimation accuracy and computational efficiency than extended Kalman filter (EKF), unscented Kalman filter (UKF) and traditional CKF while meeting real-time constraints.

Abstract:This paper proposes a structure-inspired parameter estimation method for the composite load model with distributed generation (CMPLDWG) developed by Western Electricity Coordinating Council (WECC). The high dimensionality and strong nonlinearity of this model, due to the aggregated distributed energy resource (DER_A) component, greatly complicate the reliable parameter estimation. We put forth a parameter interdependency analysis to partition the entire parameter space into smaller subsets, thereby decomposing the original high-dimensional estimation problem into multiple tractable subproblems. After applying the interdependency-based parameter grouping, the estimation for each subset is performed using both the Levenberg-Marquardt (LM) algorithm and the enhanced snake optimizer (ESO), demonstrating the solver-agnostic improvements in the convergence stability and estimation accuracy. An initialization strategy is developed to improve the robustness of subsequent optimization. Case studies in the New England 68-bus system confirm that the interdependency-based parameter grouping significantly improves the convergence speed, numerical stability, and estimation accuracy across different disturbance scenarios. The ability of the proposed estimation method is further validated through real-world measurements, and it can be broadly applicable to other modeling problems.

Abstract:The transition to sustainable energy systems demands robust and reliable control strategies to ensure stability and performance under dynamic and uncertain conditions. This paper presents an integral quadratic constraint (IQC) based framework for robust control synthesis and analysis in a microgrid, with guarantees of robust stability and performance. The IQC theory offers a powerful mathematical framework for analyzing the impact of uncertainties and nonlinearities in energy systems, providing more generalized and flexible tools compared to the widely used structured singular value i.e., μ-synthesis. The application in the microgrid control system synthesis demonstrates the superior performance and robustness of the proposed IQC-based framework in the presence of parameter variations and exogenous disturbances. This paper bridges the gap between advanced control theory and its practical applications in microgrids, offering a new perspective for researchers and stakeholders seeking innovative solutions for the control system design and analysis.

Abstract:The synergistic installation of energy storage (ES) and photovoltaic (PV) systems in industrial microgrids plays an irreplaceable role in improving energy efficiency, curbing carbon emission growth, and promoting sustainable economic development. However, the high initial investment costs of ES, coupled with the inherent uncertainty and volatility of PV, have restricted the large-scale application of ES. To address these challenges, this paper proposes an enhanced bi-layer iterative stochastic robust planning method for shared rental ES (SRES) in industrial microgrids. The upper layer develops a multi-objective probabilistic information gap decision theory (IGDT) method to investigate the optimal capacity and power of SRES among industrial microgrids under PV uncertainty. The lower layer proposes a demand power defense-driven distributed model predictive control (DMPC) method to manage the leased power from SRES to industrial microgrid. Numerical results demonstrate that compared with the self-built ES and shared ES, SRES achieves economic benefit improvements of 2.37% and 2.42%, respectively. The proposed planning method effectively mitigates the impact of PV uncertainty, enhances demand power defense capability, and significantly improves overall economic efficiency.

Abstract:Low-voltage photovoltaic (PV) expansion reduces energy revenue of utility grid but leaves reserve capacity and costs unchanged and hard to quantify. To address this issue, the utility grid promotes the green power direct supply scheme that deliver PV energy to consumers in behind-the-meter (BTM) community without relying on public transmission lines. With a clear ownership boundary, the utility grid can accurately measure and share the previously neglected cost of providing reserve capacity, thereby encouraging PV power generation to seek local reserve capacity in the BTM community to reduce cost. Dedicated slow-charging electric vehicle (EV) facilities in the BTM community, servicing EVs that have fixed arrival and departure schedules and long parking durations, can offer local reserve capacity to mitigate PV power deviations. This paper proposes an EV incentivization framework to provide local reserve capacity for PV power generation in the green power direct supply scheme. First, a stochastic optimization model is established to evaluate expected residual PV power deviation costs and determine the corresponding local reserve capacity demand. Second, a bi-level Stackelberg game model is formulated to describe the multi-round bargaining between the charging facility operator (CFO) and EV users. Simulation results demonstrate that the proposed EV incentivization framework effectively aggregates EV adjustable power, reduces PV power deviation penalties, and alleviates the reserve capacity burden of the utility grid.

Abstract:Peer-to-peer (P2P) energy and carbon sharing among prosumers promotes the local decarbonization, yet it faces new challenges due to the dynamic roles and heterogeneous individual characteristics of prosumers. This paper considers the social behavior and low-carbon preferences of prosumers, and proposes a novel matching-based energy and carbon sharing scheme. First, a bidirectional carbon emission obligation transfer model via sensitivity coefficients is developed, enabling dynamic allocation of emission obligation based on the energy sharing strategies of prosumers. Second, the hybrid preference model of prosumers in P2P energy and carbon sharing is formulated, which integrates, besides economic-based preferences, the considerations of social behavior and low-carbon preferences. Third, a supply-demand ratio (SDR) based proposal method combined with a modified stable matching algorithm is developed, achieving faster market clearing than conventional alternating direction method of multipliers (ADMM) methods while guaranteeing weak Pareto-optimality. Implemented on an IEEE 33-bus system with 32 prosumers, the proposed matching-based energy and carbon sharing scheme can enhance the participation of prosumers, encourage prosumers to choose cleaner energy, and then reduce the total carbon emissions of cluster.

Abstract:Power systems with renewable energy penetration increasingly rely on flexible loads to provide ancillary services, including emergency DR renewable energy consumption, peak shaving, and valley filling. However, the demand response (DR) characteristics of diversified loads such as long DR delay time and ramp-up time do not match the high-resolution bidding and fast-response requirements of ancillary services. To address this challenge, this paper proposes a multi-market DR bidding strategy for load aggregators (LAs). First, the specific bidding resolution and response frequency rules of emergency DR, renewable energy consumption, peak shaving, and valley filling are analyzed, leading to the establishment of a market-side time granularity model. Meanwhile, the DR characteristics of flexible loads, including DR delay time, ramp-up time, and duration, are reformulated into a load-side time granularity model based on DR deviation penalty rules. A bilateral time granularity matching framework is then constructed to enable coarse-grained loads to participate in fine-grained markets through coordinated aggregation. To enable the distributed optimization of DR bidding plans and load response schedules under the management of LAs, a multi-market DR bidding strategy based on bilateral time granularity matching framework is developed. Case studies demonstrate that the proposed strategy effectively unlocks the time flexibility of diversified loads, allowing them to participate in multiple markets and maximize economic benefits.

Abstract:As an emerging demand response (DR) resource, data centers (DCs) have garnered significant attention due to their ability to adapt to stochastic renewable energy fluctuations. To fully exploit the price signals from diverse markets and achieve the flexible complementarity between DCs and virtual power plant (VPP) resources, this paper proposes a coordinated framework for the data center virtual power plant (DCVPP) participating in the electricity -carbon joint market under uncertainty. Specifically, a refined model of power consumption for cooling system in DC is developed based on the second-order equivalent thermal parameter (ETP) theory, which enhances the accuracy of load modeling and maximizes the potential for DR. To address the penalty risks associated with forecast uncertainties in renewable energy generation and load demands, a risk-based reserve scheme is developed by integrating the Gaussian copula with two-sided superquantile theory. This scheme provides a coordinated mechanism for quantifying the correlations between forecast errors. On this basis, an optimal scheduling strategy of DCVPPs is formulated in the electricity-carbon joint market, which coordinates DC load flexibility with distributed energy resources to optimize economic performance and reduce carbon emissions. Case studies demonstrate that the proposed refined model achieves an root mean square error (RMSE) reduction ranging from 57.9% to 68.8%, and the risk-based reserve scheme reduces penalty costs by 24.4%. Furthermore, the integration of the carbon trading mechanism in the proposed optimal scheduling strategy results in an increase of $259.9 in economic benefits and an average reduction of 14% in carbon emissions.

Abstract:The increasing electricity demand imposes higher requirements on the power system to generate electricity and balance the supply and demand. The unit commitment problem is critical in power systems for determining the generation schedules and market prices. However, its inherent nonconvexity and discontinuity pose significant challenges to efficient pricing. The convex hull method addresses this nonconvexity by taking the slope of the convex envelope of cost function for generating units over the convex hull of the feasible set as the price. Nevertheless, due to the nonsmoothness, the convex hull price cannot be computed by using the gradient-based methods, which increases the computational complexity. Smoothing techniques provide an effective way by converting the nonsmooth problem into a smooth one, which enhances the computational efficiency. This paper proposes a real-time pricing scheme for the smart grid that incorporates renewable energy generation and energy storage devices on the demand side and multiple generating units on the supply side. Considering the dynamic demand of users and startup costs of generating units, a social welfare maximization model is formulated. By employing the convex hull method, the convex envelopes of the cost functions for generating units are derived. The smoothing technique is then employed to convert the original model into a smooth one. Furthermore, a distributed iterative algorithm on the basis of gradient projection method is developed by leveraging the separable structure of the variables to solve the model efficiently. Simulation results validate the feasibility and effectiveness of the convex hull method for real-time pricing in smart grid with multiple generating units via smoothing technique.

Abstract:Line-commutated converter based high-voltage direct current (LCC-HVDC) systems may diminish the power grid strength, potentially precipitating static voltage stability issues. To maintain the required level of power grid strength while minimizing the total system cost, this paper proposes a coordinated expansion planning framework for multi-infeed LCC-HVDC (MI-HVDC) systems that explicitly incorporates generalized short-circuit ratio (gSCR) constraints. Specifically, the gSCR requirement is innovatively formulated as a semidefinite constraint, and the proposed framework is reformulated as a mixed-integer semidefinite programming (MISDP) model to guarantee the global optimality. Then, to address the computational intractability of the MISDP model, a generalized Benders decomposition (GBD)-based approach is employed for efficient solution. Time-domain simulations conducted in PSCAD/EMTDC demonstrate the effectiveness of the proposed framework in enhancing the voltage stability of MI-HVDC systems.